Chapter 1 The temporal-spatial parameters ee am a slow walker, but I never walk backwards. ‘Abraham Lincoln CHAPTER CONTENTS The gait cycle: steps strides 16 Effect of speed 24 Key events in the gait cycle 16 Effect of speed on the temporal phases of gait 25 Double support 17 Effect of age: maturation 25 Variability of the TSPs 29 a Walking speed 18 Cadence 19 Stride length 20 Uses of the TSPs 21 Treadmill gi Control of the TSPs 31 Normative values 23 Upper-limb motion 34 OBJECTIVES ‘© Learn the basic terminology of the gait cycle ‘© Define the temporal-spatial parameters of gait and be able to measure them clinically ‘© Awareness of the equipment available for measurement of the temporal-spatial parameters ‘© Appreciation of the effects of age and speed on the temporal-spatial parameters ‘© Appreciation of the use of normalization in reducing the variation caused by stature ‘© Know how to interpret the temporal-spatial parameters in healthy and patholagical gaitPart | THEORY THE GAIT CYCLE: STEPS & STRIDES Each time a leg goes forward, it makes a step. For example, when the right leg goes forward, it makes a right step, when the left swings forward it makes a left step. Step length is the distance from the heel of the trailing limb to the heel of the leading one, ‘When one of each (right and left) has occurred, the person has taken a stride, or performed one gail cycle, and the time it takes for this to occur is called the gait cycle duration, or stride time (Fig. 1.1), KEY EVENTS IN THE GAIT CYCLE Figure 1.1. One gait cycle, or stride, is completed when two steps have been taken, Stride Tength is usually measured ‘between two successive contacts of the same foat. It’s usual to start the cycle (0%) with the first contact (initial contact, often. called heel contact in normal gait) of one foot, so that the end of the cycle (200%) occurs with the next contact of the same (ipsilateral) foot, which will be the initial contact of the next cycle. In normal symmetrical walking, toe-off occurs at about 60-62%, dividing the cycle into stance (when the foot is on the ground) and swing phases (Fig. 1.2). Since there are two lower-limbs, the events on the opposite (contralateral) limb are offset by 50%, so contralateral initial contact occurs at 50% cycle ‘When one limb is in swing phase, the other is in stance (Fig. 13) Lett step Rihtstep One sideChapter 1 The temporal-spatial parameters Figure 1.2 Toe-off divides the gait cycle into stance and swing phases, 0 2 0 o ® 00 Inia Tort bial contact ceiact Figure 1.3 Left intial contact loceurs when the right side is at 50 cycle, Ipsilateral intl Totot tania DOUBLE SUPPORT Since each stance phase is 60%, and 2 x 60 = 120, it follows that for 20% of the cycle both feet are on the ground. This time period is called double support, and is, in fact, the definition of walking: as speed increases, double support time falls (Kirtley et al 1985), and running begins when it becomes 0% (i. when stance duration is 50%). The double support time is divided into two parts, which can be termed initial (in which weight is being transferred from contralateral to ipsilateral) and terminal (in which weight is being transferred from ipsilateral to contralateral limb). Of course, the initial double support of one limb is the same as the terminal double support of the opposite one (Fig. 14), Knowing the stance duration, the double support can be calculated, and vice versa: Total double support time, DS = Stance ~ Swing. But, Swing = 100% —Stance, so DS = Stance - (100 - Stance) ive, Stance = (DS + 100)/2 v7 —18 Part | THEORY = Figure 1.4 There are two ni Terminal petiods of overlap (double double support double support support), when both feet are on the ground tia Teeat tial cont coniat ‘The part of stance phase between the double support phases (when only one foot is on the ground) is called single limb support (SLS). | Mca 1.1 What would be the stance duration for a gait with a total double support of 30%6? (50% (b) 60% (©) 65% (@) 70% WALKING SPEED Walking speed can be calculated from the equation: Speed = Distance/ Time Although everyone has a natural (free or self-selected) walking speed, the actual speed is continuously adjusted according to the conditions. Speed must be quickly slowed or increased to avoid collisions with other pedestrians or vehicles, and is consciously or subconsciously varied according to mood and schedule, When two people walk together, they rapidly adopt a mutually acceptable speed in order to walk together. It almost goes without saying that most walking problems result in a reduction of speed. [P| mea 12 sos ang would i tte to wal 30m at a peed of 15 m2 (5s (b) 10s () 15s (@) 20s Since there is no mechanical coupling between the lower-limbs, the coordination responsible for the alternating motion ofthe legs must arise somewhere within the nervous system. In eats, thythmic locomator patterns are generated in the spinal cord by self-sustaining circuits called central pattern generators (CPG). As every parent knows, newborn babies will make stepping movements when their feet are touched alternately onto aFigure 1.5 (Reproduced by permission of Dr J Hider, National Rehabilitation Hospital, Washington, DC) Chapter 1 The temporal-spatial parameters surface. Similarly cats will walk on a motorized treadmill even if the whole brain is removed (or if the thoracic spinal cord is transected). This observation has given rise to a new approach to gait rehabilitation called body weight support therapy. The appropriate sensory feedback needs to be synchronizes to events in the gait cycle, eg. when one limb is loaded the contralateral limb must be unloaded. Robotic exoskeletons (Fic. 1.5), such as Lokomat™ (Hocoma AG, Zirich, Switzerland), have been devised to automate the therapy, which is now being used for patients who have had a spinal cord injury (SCI) or stroke. I 6 CS 7 J en Pa [) \ : a Dieta V 2008 Spinal cord pattern generators fr locomotion. Clinieal Nevtophysilogy 4(g):1379-1388 JezemicS, SeharerB, Clams G, MorariM 2008 Adaptive rbot rehabilitation of locomotion: 2 cial study in spinaly injured ivi, Spinal Cor 41857668 19 — CADENCE ‘The number of steps per minute is called the cadence. Really, it would bbe more logical to express this as strides (or gait cycles) per second. However, cadence is an old concept that originated in the military, and steps are easier to count than strides (try it). Since there are two steps (left and right) in every stride, and 60 seconds in one minute, steps per minute can be converted to strides per second by dividing by 120. Note also that the stride time is simply 120/Cadence. Natural cadence is a little less than 120 steps/minute (i. about one gait cycle per sec- ond). Similarly, the cadence (in strides/s) is the reciprocal of the stride time, s0 a cadence of 120 steps/min is equivalent to a stride time of 120/120 = 1s20 = Part | THEORY 2 Mca 1.3 Figure 1.6 Cadence decreases 25 limb length inereases, in a similar way to the pendulum of 2 clock (data from Sutherland 1994), STRIDE LENGTH What is a cadence of 80 steps per minute expressed in stdes/s? @o3 Wos (©) 087 @os Cadence is related to the length of the lower-limb in a similar fashion toa pendulum in a grandfather clock: longer legs have a slower cadence (Fig. 1.6). Consequently, most people seem to maintain a constant ‘walk. ratio’ (stride length divided by cadence) throughout life (Sekiya & Nagasaki 1998). Since women are, on average, alittle shorter than men, they tend to have a slightly higher cadence. Small children have an even. more rapid cadence (up to 180 steps/min), which gradually falls as they {grow taller (Rose-Jacobs 1983). Cedence=-95 Loyertinb length + 200 aoa oso 035 0d oS 8085s Lowers length (n) ‘Walking speed is the product of cadence and stride length. Since cadence is usually measured in steps/minute, it needs to be divided by 120, so the final equation is Speed = (Cadence x Stride Length)/120 This is a very useful equation, because it enables any one of the three variables to be calculated, given the other two. Speed is easily measured from the time taken to walk a known distance, and by watching carefully the number of steps taken in that time can be counted, Stride length can be determined by dividing the distance travelled by the number of steps and doubling the result (I stride = 2 steps), or by rearranging the above equation: Stride Length = (120 x Speed) /Cadence eg. given a cadence of 120 steps/min and walking speed = 1.5 m/s, we can calculate:Chapter 1 The temporal-spatial parameters 21 Stride Length = (120 x 1.5)/120= 1.5 m Since stride time is the reciprocal of cadence (in strides per second), the equation can also be written: Stride Length = Speed x Stride Time Note that, like walking speed, stride length has to be constantly adjusted in order to negotiate kerbs, potholes, corners, steps, etc. The dependence of speed on both variables allows a flexible range of combi- nations of cadence and stride length to be used to maintain speed under a variety of circumstances. Together the walking sped, cadence and stride length are calle the temporal-spatial parameters (TSP) of gait, and their measurement forms the basis of any gait assessment. 2] Mca 1.4 What is the stride Jength of a person walking at 1 m/s and 120 steps/min? (2) 0.5m (0) 1m (12m @15m 2] Mca 1.5 What is the stride time of a person with stride length 1.2 m and speed 08 m/s? @)oss (oars (125 (iss USES OF THE TSPs ‘The TSPs are important functional measures — the vital signs of gait. The main applications for them are «© Screening (e.g. to detect elderly people at risk of falling) @ Asa performance measure (e.g. to grade a patient's level of disability) '® Monitoring the efficacy of therapy (i.e. a an outcome measure) © Normalization of other gait measurements (in order to compare results from people walking at different speeds) CS CLINICAL POINTER ~ MEASUREMENT OF THE TSPs [Although a stopwatch and a measured distance are all that is needed to measure the TPs, there are also several commercial systems available for automated measurement, The simplest of these is an optical or infrared detector (eg. the Speedlight Timing System, SWIFT Performance Equipment, Alstonville, Australia, cost US$3,380; Powertimer, Newtest, Oulu, Finlané, cost US$4300), which turns 2 timer on and off when the subject breaks two light beams placed a known distance apart (Mitchell & Sanders 2000}. Ths is fine for measuring walking speed, but an observer is22 Part | THEORY Figure 1.7 Nike sdm ftriae 100 Nike, Portia, OR). The wristwatch cleulates the wearers speed from rao signals receives trom 2 sensor in the shoe pod. Figure 1.8 Stride Analyzer footswitch-based portable gait analysis system (B & L Engineering, Tustin, CA) still needed to count the numberof steps taken in order to calculate cadence and stride length To count the number of steps, a pedometer can be used. These are inex- pensive devices (ca. US$20) that can be attached to the subject's belt, and they count each time they detect a step (Sequeira et al 1995). Unfortunately, they are prone to over and underestimating (Schmalziee ct al 1998), especially in women, where the errors can rise to 24% Silva et al 2002). Some have an activation threshold that can be adjusted ta try to prevent this but it tends to be subject-specific. Thus, pedometers are only really useful for measuring the approximate number of steps taken over prolonged periods, ie. activity monitoring. New accelerometer and gyro sensors offer the possiilty of long-term ambulatory monitoring (Coleman ‘et al 1999, Aminian et al 2002, Macko et al 2002, Kirtley 2002), and several products using these sensors are now available: eg, FitSense FS-1(FtSense Technology Inc., Southborough, MA), and Nike sdm|[triax 100] (Nike, Portland, OR). In both cases, a wristwatch calculates the TSPs after receiving racio signals from a small pod attached to the subjects shoe (Fig. 1.7). They are mainly aimed at runners to track their workout but can be useful clinically, and the data can be recorded on the watch and downloaded later to a computer. Ultrasound ean also be used to measure the speed of a walking person, in 2 similar fashion to police radar guns (Huitema et al 2002). Small fluctuations in the recorded speed can then be used to determine cadence, and compute stride length. Global Positioning Satellites have even been used (Terier et al 2000) to measure waking speed outdoors. The temporal parameters of gait (stance, swing and double support times) are more dificult to measure accurately Footswitches (Hausdorff et al 1995, Blanc et al 1999) on the toe and heel can be used. They make an electrical contact when that part of the foot is loaded, e. Stride Analyzer (8 & L Engineering, Tustin, CA; cost USS9,900; Fig. 1.8) calculates velocity, cadence, stride length, the duration of single and double support for each limb, and the pattern of contact for each foot (Hill etal 1994, Goldie et al 1986) Alternatively, video can be used together with a time code generator (Wall & Crosbie 1996). (One method that has recently become very popular uses a special instra~ mented walkway, €.. GaitMat !* (E.0. Inc, Chalfont, PA; cost USS14,500) and GAITRIte™ (CIR Systems In, Clifton, NJ; cost US$14,000; Fig. 1.9). The GaitMat I! uses an array of 38 rows x 256 switches to record each footfallChapter 1 The temporal-spatial parameters 23 Figure 1.9 GAIIRite pressure sensor array instrumented carpet (CIR Systems Inca, Clifton, NJ} (aiving 2 spatial resolution of 15 mm), with dimensions of 384 x 60 x3 em high in three hinged sections. GATRite™ uses a pressure-sensing array arranged in 48 rows of 288 sensors to record the imprint of each footfall with six different levels of pressure (McDonough et al 2001), Itis only 3 mm thick, ‘s0 can be rolled up and transported, The most common mat, 4 m long, weighs 20 kg in its case, but a 7 m version is also available. NORMATIVE VALUES It has to be said that despite all the technology that is available and the simplicity of measuring the TSPs with a stopwatch, they are, in fact, rarely measured in routine clinical practice. One possible reason for this is that although several studies have reported normative values, there is unfortunately no good consensus about what the normative ranges should be. A person’s natural gait is very dependent on the environment, for example, people tend to walk faster on a long walkway (Murray et al 1966, 1968, 1970), and slower on a short one (Oberg et al 1993), and are also influenced by the size of the room. Consequently, outdoor studies (eg. Finley & Cody 1970, Waters et al 1988, Hausdorff et al 1999) invari- ably report higher speeds and stride lengths than indoor studies (Grieve & Gear 1966, Oberg et al 1993). This is rather frustrating, and means that normal values should really be obtained for each laboratory or clinic. Bearing this in mind, Table 1.1 is a rough guide, given a typical (ca. 5m) ‘walkway length. I's easy to remember 1.5 m/s, 15 m and 120 steps/min (ie, one stride/s) as a quick rule of thumb. aomnuine ingest be TS Speed (mis) Cadence Stepsimin) Stride length (m) Men 1316 0-115, rer Women 12-18 1-120 13-1524 Part | THEORY EFFECT OF SPEED Figure 1.10 Cadence nereases linearly with speed, Figure 1.11. Stride length also mereases with speed, but the relationship is logarithmic rather than linear Walking speed is related to both cadence and the stride length, so it can be increased by a more rapid cadence, longer stride length, or both. In. healthy people (those with no gait disorder) both parameters increase with speed. Cadence increases linearly (Fig. 1.10) but stride length increases logarithmically (Fig. 1.11), changing a lot at low speeds, but tending to level off at higher speeds, 00 10 10 wo 1% a g j 100 ob 10 15 20 28 30 length (m) Side eng = 0.88 op(Speed) + 1.18 P08 88 oo 08 10 18 20 25 Speed is)Chapter 1 The temporal-spatial parameters EFFECT OF SPEED ON THE TEMPORAL PHASES OF GAIT Table 1.2. Relationship between speed and the ‘temporal parameters | Mca 1.6 As speed increases the double support time (along with stance duration) decreases. When double support reaches zero, running begins, and with further speed increase, the double support phase becomes negative (ic. it becomes a fight phase). Some useful relationships (Blanc et al 1999) are given in Table 1.2. Variable Men ‘Women ‘Stance duration (96) 71 x stride time (s)— 11.371 x stride time — 10.9 Double support (98) 41 x stride time (5) 203 41 x stride time — 20.0 CLINICAL POINTER ~ BALANCE COMPENSATION Stance phase is also slightly longer while walking in bare feet compared to when wearing shoes (Eisenhardt et al 1996), Shoes provide a slightly increased base of support, which helps balance, As balance is compromised, both stance and double support increase to provide an increased support time. This is an example of a compensation strategy. Which gait would you expect to have the longest double support? (2) Barefoot at 1 mis (6) With shoes at 1 m/s (€) Barefoot at 1.5 m/s (4) With shoes at 1.5 m/s EFFECT OF AGE: MATURATION ‘The average onset of walking in children is about 11 months (mean 329 + 46 days), and, interestingly, baby-walkers seem to delay this by about 21 days (Garrett et al 2002). Using a normative range of Mean + 3 SD, this ‘would suggest that a baby should walk before around (329 +346) = 467 days (i.e. 16 months). In practice, parents are usually reassured up until the child is 18 months old (Sutherland et al 1980, 1988). To determine the age at which gait matures (ie. becomes adult-like) we need to remove the effect of size since children are smaller than adults (Sutherland 1996). This technique is called normalization. Probably the best (but not always used!) method involves convert- ing gait variables into dimensionless quantities (Hof 1996, Sutherland 1996, Pierrynowski & Galea 2001, Stansfield et al 2003). For example, normalized stride or step length is obtained by dividing by lower- limb length, I, measured from anterior superior iliac spine (ASIS) to medial malleolus (Fig. 1.12). Walking speed must be divided by V/gl) where g is the acceleration due to gravity (9.81 m/s"), whereas cadence is divided by (g/l). When this is done, it appears that the 25 —26 Part | THEORY Figure 1.12. Lower-imo length is measured from ASIS to medial malleolus, Figure 1.13 After normalization, walking speed is seen to stabilize at around 4 years of age (data from Sutherland 1984). oer sper | Tac pe Meal males temporal-spatial parameters stabilize at around 4-5 years of age (Figs 1.13, 1.14, 1.15), An alternative approach to normalization involves plotting the TSPs fon a nomogran, such as those developed by the Shriners groups of hospitals (Todd et al 1989, Johanson et al 1994) In the clinic, a useful rule of thumb to remember is that a child’s natural stride length should be about 90% of their height (Fig. 1.16). ome ‘Age (eats)Figure 1.14 Normalized 1s stride length also stabilizes by 10 ‘around 4 years of age (data from Suteind 1804) 158 150 Hive is 2 18 2 Figure 1.18 Normalized cadence stabilizes by around § years of age (ata from Sutherland 1994), Figure 1.16 An approximate guide to normal stride length in children is 80% of height (Gata from Hausdorff etal 1999) Age years) ‘ge years) Height)28 Part | THEORY Figure 1.17 Natural walking speed remains fairly constant, bout fast speed declines after age 50 years (data from Oberg et al 1993). Remember Menelaus's ‘rule of thumb Total adult height jeight at age 2y x2 The distal femoral epiphysis contributes 10 mm/y (70% of femur growth), while the proximal tibial epiphysis generates 6 mm/y (60% of tibia growth), Growth ceases at 14-15 years in girls and 16-17 years in boys, CLINICAL POINTER - DEGENERATION OF GAIT Natural walking speed remains relatively stable until about age 70 (Winter et al 1990, Leiper & Craik 1981); it then declines about 159% per decade (Fig. 1.17). Healthy subjects can increase their speed by as much as 442% above natural pace (Finley et al 1969), However, maximal speed declines earlier and more steeply: about 20% per decade after the age of 50. Cadence does not change with age (maintaining its relationship to lower limb length}, so stride length must be the source of the decreased speed, ‘One place where walking speed really matters isthe pedestrian crossing, Most crossings are designed for a walking speed of around 1.2 m/s, meaning that about 15% of elderly pedestrians have difficulty getting across before the lights change (Coffin & Morrall 1998). ‘As people age, balance slowly deteriorates, and this i reflected in the temporal parameters, Stance accounts for 58% of gait cycle at age 20, and 66349 at age 70, with double support duration increasing from 18% to 26% (Murray et al 1969), Interestingly, reduced stride length, reduced speed and increased double support time seem to be associated with fear of falling, rather than falling itself (Maki 1997) 20 ‘Age (ers Biometries is the science of recognizing people by their biological attributes, Most biometrics applications require the subjects cooperation, e.g, people must put their finger onto a scanner to read their fingerprint. Video surve lance cameras can monitor people without their knowledge, but they require human operators to monitor the images for suspicious events. In an effort to automate the pracess, attention has therefore turned to gait, and especiallyChapter 1 The temporal-spatial parameters 29 the ratio of stride (or step) length to cadence (the so-called waik ratio) (Fig. 1.18). This may be a signature that may discriminate between, e., humans and non-humans, adults and children, men and women, walking and running, ete, and it may eventually be possible to detect aggressive or malevolent behaviour. Watch out - Big Brother may be analysing your gait! Figure 1.18 Relationship between individual stride length and cadence, (Reproduced by permission of BenAbdelkader et al 2002) ‘Stride length (m) ® 0 100 10 120 10 10 Cadence (stepsimin) ‘BenAbelkader C, Cutler R, Davis L 2002 Stride and eaence 25 a biome in automatic person dentfieation and vr‘ieation. Proceedings of te Intemational Conterence on Pattern Recognition, 11-15 August, Gutbec City, Canada VARIABILITY OF THE TSPs As with all gait variables, itis important to bear in mind the natural bio- logical variability that is invariably observed when a measurement is repeated several times. Variation can occur within the same subject (intra-subject), oF in a group of subjects (inter-subject). Not surprisingly, the latter is usually higher, because people walk differently. The amount of variation is also usually less when the measurements are taken at the same time (intra-session) versus a different time of day or different day (inier-session variation). Since there is bound to be some measurement error, the variation recorded will also depend to some extent on the method (e.g, instructions given, environment, length of walkway, etc), ‘operator (e.g. whether the same person takes the measurement, and how skilled the person is) and instrument (stopwatch or some more sophisti- cated method) used.30 Part | THEORY Figure 1.19 Gait is more variable in the early years, but gradually stabilizes with maturity (see Figs 1.13-1.15). The CV of stride time falls from around 6% in 3-to 4-year-olds, to around 2% in 11- to Ld-year-olds (Haussdorff et al 1999), In the elderly, increased variability seems to be associated with risk of falling (Hausdorff et al 1997a,b), with speed variability the single best predictor of falls (Maki 1997) Gait cycle duration is rarely constant, and until recently these step-to-step fluctuations were just assumed to be random noise superimposed upon 2 constant walking rhythm, Another possibility is that there might be some underlying complex temporal structure to the variation, Le, non-linear, or fractal phenomena. If this were the case, there would be a ‘memory effect the stride time at any instant would depend on previous stride times. Fractal ‘behaviour, such as has already been observed in long duration heart rate recordings from 24-hour Holter monitoring, may confer important biological advantages related to adaptability Long-term measurement of stride time has been done using in-shoe footswitches (Peng et al 1999), A process known as Detrended Fluctuation Analysis (OFA) was then used, Ths isa bit complicated, but basically involves dividing the time series (gait cycle duration plotted against time) into boxes of equal lenath, nn each box a straight line is fitted to the data (represent ing the trend in that box), finaly resulting in a quantity called Fl The relationship between fog Fn) to fog nis usually found to be a straight line The slope of this line, ct has been determined to be O.5 for random noise, 1 Orghaldsta a= 08s © Shutled data a= 05Chapter 1 The temporal-spatial parameters 31 with higher values indicative of long-range correlations (memory) in the data, It turns out that a = 0.83 for the gait data, suggesting that there is indeed some pattern to the seemingly random variation in stride time. When the data are randomly shuffled to remove any memory effect, falls to 0.5, as expected (Fig 1.19). Its possible that there may be clinical significance in this fining, because values of o were found to be closer to 0.5 (more random) for elderly people compared to young subjects, and in patients with Parkinson's disease and Huntington's chorea (Hausdorff etal 1998). Peng CK Hausdorf{J M, Goliberger AL 1999 Fractal mecharisms in neural contol: human heartbeat and gait yams in neath and dsase, I: Walezek.J (a) Nonlinear dynamics, ‘el-organiation, and biomedicine. Cambiiage Universty Press, Camoridge ausdrt# JM, Cudkow’ez ME, Firion Re al 199 Gait varity and basal ganglia disorders: stride-to-strie variation of gait ele timing in Parkinsons and Huntingtons esease Movement Disorders 13:428-437 TREADMILL GAIT Many times in rehabilitation, subjects are asked to walk on a treadmill = the advantages being safety (because a harness and/or handrails can bbe used), less space is required, and walking speed can be directly con- trolled. Gait on a treadmill is not quite the same as free walking, however. In particular, the stride length on a treadmill (or more correctly, contact or support length) is shorter and cadence higher for a given walking speed (White et al 1998, Alton et al 1998). Contact length = 0.665 + 0.25 x belt speed, so for a speed of 1.5 m/s, contact length would be only 1.04 m (Kram & Powell 1989). Conversely, cadence increases by 7% in adults and 10% in children compared to overground walking, while stance phase duration decreases by 5% in adults but curiously remains unchanged in childzen (Stolze et al 1997). DEBATING POINT Debate the advantages and disadvantages of making gait measurements on: = Overground waking = Treadmill gait CONTROL OF THE TSPs Since both stride length and cadence are under voluntary control, they can be manipulated by external cues (Zijlstra et al 1995). Subjects will, for example, synchronize their cadence to the rhythmic sound from a metronome (frequency modulation, FM), or adjust their step Tength to lines drawn on the walkway (amplitude modulation, AM). Interestingly, the normal temporal parameters of gait (stance, swing and double support percentages) are maintained (invariant) during FM bbut not AM32__Part | THEORY ‘More information comes from experiments with a split-belt treadmill in which the left and right legs move at different speeds (Zijlstra & Dietz, 11995), Although this sounds impossible, subjects are easily able to auto- matically adapt their walking pattern after only a few seconds even when cone belt moves at four times the speed of the other. Stance phase duration is found to increase on the slower belt and decrease on the faster side in order to keep the contact lengths equal on each limb. Presumably, afferent feedback (perhaps from pressure receptors on the sole of the foot, or joint spray endings) via the spinal cord is responsible for this adaptation, ‘These experiments suggest that a central pattern generator (CPG) provides the basic stepping rhythm, but its rhythm is modified by feed- back according to the biomechanical context. CLINICAL POINTER ~ POINTER-INTERPRETATION OF Ci THE TSPs IN DISEASE Walking speed is easily measured and has been shown to correlate well with function. Sometimes walking speed alone can be misleading, since it is @ product of cadence and stride length, Most gait problems result in a decreased stride length, and so an increased cadence is a comman compen- sation to maintain speed. In some diseases (eg. the festinating gait of Parkinson's disease), the increase in cadence is very marked (Morris et al 1996, 1998). It is therefore best to measure all three parameters Table 1.3). AAs mentioned previously, stance duration is often increased when balance is compromised (dysequilibrivm) due to vestibular, cerebellar (ataxia) or nan- specific instability. Conversely, it is decreased ifthe leg or arasthesis on that Side is unstable or painful (antalgic gait), and the contralateral stance dura tion will be prolonged to compensate (Table 1.) Table 1.3 Interpretation of 5 ispsin gait disorders [S664 Stride length Cadence Conel N N N Normal N L t Compensated gait 4 J t inadequately compensated gait u J b Severe gat impairment Table 1.4 Interpretation of ‘temporal measures in gait | Stance duration Conelusion| disorders L Ipsilateral pain or instability. t Dysequilbrium or contralateral instability What is the most ikely diagnosis in a patient with stance duration = 55%? 2| Mca 1.7 (@) Ataxia (b) Antalgic gait (©) Balance problem @) childFigure 1.20 Step width is increased in disorders of balance in order to increase the base of support Figure 1.21 In this gait, right step length is shorter than the left, but note that the stride lengths are still equal Chapter 1 The temporal-spat parameters ‘The step width (mediolateral distance between the heels in double sup- port) also tends to increase with dysequilibrium in order to increase the tase of support (Fig, 1.20), although this may only become evident at higher speeds (Krebs et al 2002). Normal step width varies with age, so it’s best to normalize it by dividing by pelvic width (measured between the ASIS): normal < 05, ie. less than half the pelvic width. Interpretation of step lengtit sometimes causes confusion. Since the legs are connected together at the pelvis, both legs must travel the same dis- tance (unless the subject is walking in a circle), so the right stride length must equal the left stride length (give or take inter-stride variation). The right and left step lengths can be quite different, however, because of the way they are defined (Fig. 1.21) Tt should not be assumed that the side with the longer step length is healthier. Sometimes the sound side has a longer step length, but not always, so step length differences are useful only as a measure of sym- metry. For example, the slep length ratio (SLR) of the shorter to the longer step length is useful for tracking a patient's progress through their rehabilitation, the ratio rising closer to 1 as the gait improves. Other symmetry indices have been described (Dewar & Judge 1980, Herzog ct al 1989, Hesse et al 1999) using variables such as right and left stride times or double support times, but the SLR is probably the most widely used. These indices provide simple overall outcome measures that can even be used as biofeedback to inform the patient of their progress during gait training (Dingwell et al 1996). Ireassoo 33 —34 Part | THEORY PI Mca 1.8 UPPER-LIMB MOTION What is the SLR of a patient with right step length of $0 cm and 30 em on the left? (03 (b) 05 (06 @ 187 ‘The function of the upper-limbs during locomotion has intrigued several researchers over the years (Elftman 193%, Capozzo 1993). At natural ‘walking speed, the upper-limbs swing with the contralateral limb at a frequency equal to the cadence, whilst at slow speeds (at least in some people) they swing in phase with the ipsilateral limb at twice the cadence (Webb & Tuttle 1989, Webb et al 1994). The speed at which arm swing pattern switches from 2:1 in-phase to 1:1 anti-phase is approximately that of the natural pendulum frequency of the upper-limbs (around 05 Hz). This is an another example of a dynamic systems phenomenon, which is, typically accompanied by catastrophic flags such hysteresis and critical fluctuations that indicate instability around the transition frequency. Amplitude of arm swing increases with walking speed (Donker et al 2001), in order to counteract the angular momentum generated by the lower-limbs (Capozz0 1993, Donker et al 2002). KEY POINTS “ The gait cycle is divided into stance and swing phases % Stance phase consists of single limb support and initial and final double support phases + Walking speed can be calculated by multiplying stride length by cadence (expressed in strides/s) “ Normalization reduces the variability caused by differences in body height “ The normalized temporal-spatial parameters remain fairly constant from 5 to 70 years of age “ Gait symmetry may be assessed by the ratio between right and left step lengths or times References ‘Alton F, Baldey, Caplan S, Morrissey M C1998 A. aystem using miniature gyroscopes. Journal of kinematic comparison of overground and Biomechanics 35(5):589-699 treadmill walking. Clinical Biomechanics Blane Y, Balmer C, Landis 1, Vingerhoets F 1999 13434440 ‘Temporal parameters and patterns of the foot roll ‘Aminian K, NajafiB, Bila C etal 2002 Spatio-temporal___over during walking, normative data for healthy parameters of gait measured by an ambulatory adults. Gait & Posture 10(2)'97-108